Originally appeared in MSI Magazine
August 1, 1998

The Peace Dividend

Suppose they gave a war but nobody came? Welcome to PC Expo 98. The top stories at the show: Micron demos a 400-MHz Pentium that effortlessly displays full-screen, full motion video. Meanwhile, Sony unveils a sub-$2000 laptop that weighs barely four pounds and is less than an inch thick, while 3Com’s PalmPilot boasts an R/3-ready supply chain management add-on application.

What’s missing from this picture is the compelling need to replace corporate desktops to take advantage of the latest technologies.

In coming months, Microsoft is releasing Windows 98, a “better” Windows 95 which, guess what, does not require hardware upgrades. Over the past year, the worst shock to the system was Microsoft's Office 97. It added such “essential” features as the ability to save documents in HTML format or browse them from Internet Explorer. The price for these improvements was new file formats and APIs that were incompatible with Office 95. The result is forced migration, not so much for the new bells and whistles, but for the need to avoid a new round of compatibility hassles. But again, no new need to pump new iron.

Sometimes, there’s no choice but to keep up with the Jones. As we mentioned a couple years back (see Manufacturing Systems, July 1996), users migrate when software suppliers point a gun to their head, withdrawing support of older versions. The next moment of truth won’t occur until Microsoft readies NT 5.0, which is supposed to absorb the current Windows 9x consumer operating systems. Microsoft’s current target date for NT 5.0 is the first quarter of 1999, but don’t hold your breath. As you might recall, the original name for Windows 98 was Windows 97.

More significantly, even if Microsoft keeps to its schedule, the corporate mainstream shouldn’t have to worry about migrating to the next OS until after the turn of the century, since there is generally a 12-month lag time before a critical mass of desktop software suppliers move on to their next versions.

This is just one of many factors that are changing the very dynamics of desktop hardware asset management practices. Until now, the accepted operating life of desktop machines was assumed at three years or less. Yet, given the timeframe for phase-in of the next Microsoft OS, those 133-Mhz Pentiums that your organization bought for Windows 95 around 1995-96 should remain perfectly adequate until 2000 or possibly 2001. So much for Moore’s Law.

Of course, those five or six years may be punctuated by application software upgrades, which in some cases may force hard drive or memory upgrades, but still nothing to compare with the prospect of clearing out all those obsolete desktops.

Is it just a fluke that Intel processors have outpaced the demands of software technology, with today’s 400-Mhz screamers running Windows 9x applications only marginally better than the 133-Mhz boxes of three years ago? Was the Wintel cartel jilted by Scott MacNealy’s barbs about fat “hairball” applications, or Gartner Group’s total cost of ownership figures estimating fat clients costing upwards of $10-12,000 annually? Or, have CIO screams of “Enough, we’re not going to take this anymore?” finally being taken seriously?

Admittedly, it could be argued that Intel’'s Wired For Management and Microsoft's Zero Administration for Windows initiatives were direct responses to all TCO studies and bad press. But more importantly, while powerful desktop CPUs made the current generation of versatile, integrated desktop suites and client/server applications possible, more recent technology developments have made the powerful CPU almost irrelevant.

For instance, Internet applications are more impacted by the speed of the connection, traffic levels on the backbone, and the responsiveness of the web server, than they are by the speed of the desktop. Or the new generation of PalmPilots (and eventually Windows CE) machines, which for the first time have drawn a critical mass of applications to hand-held computers. They will work equally well connected to yesterday’s or today's Pentium boxes.

But before breaking out the champaign to celebrate the end of the (desktop) cold wars, look at the rest of your project backlogs. There are probably plenty of supply chain management initiatives, year 2000 renovations, and multi-tier application reengineering projects just waiting to claim those saved desktop dollars.